Multiple studies have shown that some patients who remain behaviorally unresponsive after severe brain damage can modulate their brain activity in response to command. A lack of response to the command- following paradigm in other patients, however, cannot be taken as evidence that consciousness is absent. There is thus an urgent need for neuroimaging tools that are able to diagnose covert consciousness in behaviorally unresponsive subjects without relying on a task. Previous neuroimaging studies of the neural correlates of consciousness (NCCs) have mostly utilized between-state paradigms (e.g., wakefulness vs. sleep) or report-based contrasts, which conflate differences in brain activity supporting consciousness with differences in brain activity supporting behavior or task response. To address these limitations, my sponsor (Dr. Giulio Tononi) has developed a within-state, no-report paradigm that uses serial awakenings to contrast states of consciousness versus unconsciousness during sleep, a physiological unresponsive state. Using high density electroencephalography (EEG), his group has shown that the occurrence of dreaming consciousness during sleep is associated with a reduction of delta activity (sleep slow waves) in a posterior cortical ?hot zone? encompassing occipital, parietal and temporal cortices. To proceed towards clinical applications, it is now essential to confirm and extend these findings using combined EEG-fMRI (functional magnetic resonance imaging). To this effect, in an important new development, Dr. Tononi and collaborators have recently shown that it is possible to faithfully track EEG slow wave activity during sleep by means of a high-frequency (~0.17 Hz) blood-oxygenation level dependent (BOLD) oscillation evident in fMRI. This finding provides an unprecedented opportunity to investigate the NCCs during sleep with high spatial resolution, avoiding potential pitfalls of EEG source modeling, and with unrestricted access to subcortical structures. Taking advantage of these new findings, in this project, I propose to combine the novel fMRI signature of sleep slow waves with a serial awakenings paradigm to locate the NCCs during sleep. Specifically, I will acquire fMRI-EEG data during ~1,080 awakenings from sleep across 40 healthy subjects and investigate the neural correlates of the presence versus absence of dreaming consciousness (Aim 1). I hypothesize that high-frequency BOLD oscillation?which tracks sleep slow-waves?will show regional decreases in specific posterior cortical areas during dreaming compared to dreamless sleep. In Aim 2, I will use univariate and multivariate modeling approaches based on high-frequency BOLD power and conventional BOLD amplitude to identify the NCCs of specific dream contents (e.g., presence of faces) and complexity. If successful, the proposed experiments will constitute an important step forward in identifying brain regions that sustain consciousness and its contents during behaviorally unresponsive states. These results can open new avenues to improve clinical care in patients with a DOC and inform the assessment of intraoperative awareness in anesthetized subjects.